Assessing the Quality of Local Government in South Africa

Literature › Works

Barak D. Hoffman (2007). Assessing the Quality of Local Government in South Africa. Center on Democracy, Development, and the Rule of Law, Stanford University (working paper). Stanford University, CDDRL.

Type	working paper
Scope	Post-1994 South African economy
Method	econometric
Period studied	1994–2005
Regions	South Africa, Johannesburg/Soweto, 243 Category B municipalities
Themes	transition post apartheid, inequality poverty unemployment, politics ideology elections, business and state, origins segregation, demography urbanisation
Citation key	hoffman2007a

Argument

After 1994 the ANC promised to expand access to local services regardless of ability to pay, but central policy now pushes municipalities toward 'cost recovery', charging market rates and often privatising services. Using municipal data for 164 or so Category B municipalities, Hoffman argues that willingness to pay (per-capita service fees) is by far the strongest predictor of core-function performance in 2005. Political competition, civil service quality and central transfers have no effect, and tests show causality runs from fees to performance. He concludes that cost recovery strengthens fiscal accountability but empties local democracy of substance and entrenches the race and wealth gaps in service access inherited from apartheid. The apartheid and transition history is brief context for why the ANC chose cost recovery.

Evidence and method

National Treasury data on service fees, taxes and transfers per capita. 2001 Census (GDP per capita, urbanisation, white share, Gini, racial fragmentation). 2001 local election results. Municipal Demarcation Board annual yes/no assessments of municipal functions (2002-2005). MDB data on civil servants' experience. Secondary sources (McDonald, Bond, Tomlinson, Makgetla, RTI 2001, SACN 2007). OLS regressions with robust t-statistics, N about 160-168.

Findings

• User fees are the dominant source of municipal revenue, so local government is largely fiscally autonomous of the centre. (p. 9)

  Moreover, user fees constitute by far the largest source of revenue for local governments, comprising on average fifty two percent of revenue.

• Apartheid local arrangements combined cheap black labour and enforced shopping patterns with residential segregation, producing unequal services. (p. 5)

  whites wanted to exploit the political inequality of blacks to subsidize white living standards through using blacks as a source of cheap labor and employing legislation to encourage them to shop at white-owned stores.

• Sanctions and instability in the 1980s caused under-investment in infrastructure, which the ANC inherited. (p. 15)

  The political instability and international sanctions that the ANC and other anti-apartheid movements catalyzed in the 1980s caused the government to under-invest in maintaining the country’s

• Fiscal constraints and the need for capital pushed the ANC to moderate its service promises. (p. 16)

  the ANC had few financial resources to meet its promise to enhance access to services, and thus needed access to capital to implement its policies.

• Under cost recovery, poor urban households bear a far heavier cost burden for the same bundle of services. (p. 13)

  poor urban households need to spend approximately thirty percent more of their income than wealthy ones to consume the same bundle of services (thirty seven percent versus three percent).

• Consumption is sharply unequal by income group, shown here for electricity. (p. 14)

  Graph four below, for example, shows that wealthy families use close to ten times more kilowatt hours of electricity than poorer ones.

• Overall local government performance of core functions is mediocre. (p. 23)

  On average, local governments undertake successfully only about fifty five percent of their core functions.

• Political competition, civil service quality and central transfers do not explain performance. (p. 24)

  political competition, the quality of the civil service (average log years of experience of the top four civil servants), and central government transfers do not affect the quality of local governments.

• Service fees are the one robust predictor of performance beyond initial performance. (p. 25)

  Third, other than initial performance, cost recovery is the only variable that has consistent and strong support.

• Other taxes do not raise performance, ruling out a general taxation-governance link or cross-subsidy from other revenue. (p. 25)

  the coefficient on service fees remains positive and significant at the one percent level, while taxes are statistically significant but have a negative sign.

• Municipalities roughly break even or slightly lose on services, so profits are not funding wider access. (p. 27)

  The median local government sells services at about a two percent loss and the average one at about a six percent one.

• The fee effect is economically large. (p. 28)

  Local governments at the 75th percentile of user fees are expected to perform thirty percent better than those at the 25th percentile.